Journal of Colloid and Interface Science 279 (2004) 458–463 www.elsevier.com/locate/jcis
In situ approaches to establish colloidal growth kinetics Lawrence D’Souza, Andreas Suchopar, Ryan M. Richards ∗ School of Engineering and Science, International University Bremen, Campus Ring 8, Bremen 28759, Germany Received 18 March 2004; accepted 27 June 2004 Available online 30 July 2004
Abstract A technique based on the back scattering phenomenon of dynamic light scattering has been employed to monitor the kinetics of gold and platinum metal nanoparticle growth and silver nanoparticle oxidation as well as in the determination of particle sizes ranging from 1 to 200 nm in diameter. The systems were chosen to examine the applicability of dynamic light scattering to nanoresearch over a broad range of sizes as well as both metallic and nonmetallic systems. The advantages of this instrumentation over traditional instruments such as X-ray diffraction (XRD), transmission electron microscopy (TEM), and scanning electron microscopy (SEM) have been highlighted. 2004 Elsevier Inc. All rights reserved. Keywords: High performance particle sizer; HPPS; Back scattering; Particle size determination; Dynamic light scattering and growth kinetics
1. Introduction Nanoscale materials represent one of the most dynamic and rapidly growing fields in modern science. While the research into these materials and their size-dependent properties is vast, there are still numerous “bottlenecks” hindering this field. One particular hindrance to the development of nanotechnology is the general lack of preparative methods that allow scientists to prepare monodisperse nanoscale materials with a predetermined size, shape and composition. The inability of scientists to observe in situ particle formation and follow growth kinetics has made this task particularly difficult. Traditional methods for characterizing nanoclusters in science and technology are very expensive and time consuming besides having their own limitations and advantages [1–4]. These methods include X-ray diffraction, transmission electron microscopy, small-angle X-ray scattering, and scanning electron microscopy [5–9]. All of these instruments have various limitations, in particular with observing particles in situ. X-ray diffraction analysis is suitable * Corresponding author. Fax: +49-4212003229.
E-mail address:
[email protected] (R.M. Richards). 0021-9797/$ – see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2004.06.086
for crystalline metallic systems with particle sizes above ≈ 5 nm, through application of Debye–Scherrer equation to the broadening of diffraction patterns for nanoclusters. Although this method can be applied to materials of smaller sizes, the results are often tentative. On the other hand, SEM can probe clusters of dimensions only above 10 nm in diameter. TEM consists of laborious instrumentation and requires very high voltages (usually about 300 kV). Furthermore, these techniques are limited to samples in powder form. TEM is only able to provide information on samples under a high vacuum and the effects of focusing an electron beam on a sample are often neglected, i.e., for many nanoscale systems the focusing of an electron beam on a small particle or group of particles can result in a rearrangement of atoms or melting. Moreover, in situ kinetics experiments are generally not possible with these instrumentations. Finke et al. [10,11] have used nanoclusters catalytic activity to follow kinetics of transition-metal nanocluster formation. This method is useful only with gas phase reducing agents but is not particularly applicable in the case of liquid phase reducing agents. Moreover, this technique is unable to provide any information about the size distribution of nanoclusters. Recently, Goodman [12] used in situ STM to study
L. D’Souza et al. / Journal of Colloid and Interface Science 279 (2004) 458–463
the mechanism, growth and thermal evolution of individual oxide-supported metal clusters. This method is quite useful to study solid state reactions and needs to develop further for other applications. One of the techniques that can overcome many of these drawbacks is dynamic light scattering [13,14], which has made significant contributions in many areas of physical chemistry. There are numerous reports available on the application of DLS in the literature. These reports are however limited to the applicability of DLS in polymer characterization and biological studies [15–19]. Recently, Jesionowski et al. [20] used DLS to study particle size distribution and agglomeration of zinc silicates; Fendler et al. [21] used DLS to characterize clay material intercalated TiO2 ; Frontera et al. [22] used DLS to study phase development in zeolites during its synthesis and Munch et al. [23] studied the aging behavior of colloidal suspensions of Laponite using DLS. Light scattering has been used to monitor agglomeration of nanoparticles and was found to be problematic due to the nonmonodispersity of particles present and the possibility of multiple scattering [24,25]. Light scattering is a powerful method for examining the structural and dynamic properties of colloids and nanoparticles, but until recently, it was not particularly effective for particles less than 5 nm in size. In general, the scattered light is collected at a particular angle; the signal shows the random fluctuations or noise. From the signals computed by autocorrelation functions, the properties of the particles are determined. The stray light equation for dynamic light scattering is given as
16π 2 n20 Rg2 2 θ 1 IS = + 2A2 C 1 + sin M 2 3λ20 4 −1 λ0 NA dn −2 GF 1 , × 2 2 I0 C 4π n0 dC −1 GF 1 1 1 + 2A2 C P (θ ) IS = , M K I0 C where IS is intensity of scattered light, P (θ ) is shape factor, M is molecular weight, C is concentration of the particles, A2 is second virial coefficient, Rg is radius of gyration, q is scattering angle, n is solvent refractive index, I0 is intensity of incident radiation, λ0 is laser wavelength, dn/dC is refractive index increment, and GF is instrument geometric factor. Recent advancement in research and development with improved electronics and measurement techniques has inspired our interest in this technique. The high performance particle sizer (HPPS) is an improved version of DLS with more advanced features. HPPS is equipped with a 3 mW He–Ne laser and a tunnel photomultiplier, both have higher sensitivity than the classical 3 W lasers and standard photomultipliers. The scattered light is collected at an angle
459
of 173◦ to the incident laser radiation, which overcomes multiple scattering, by the particles, which is minimized at 180◦ . The back scattering phenomenon minimizes the absorption of the radiation and allows measurements at high concentrations and through wide ranging particle sizes of (diameter = 0.6 to 6000 nm). The back scattering also reduces the error in measurements due to dust particles. Dust particles being large; scatter light mainly in the forward direction. Furthermore, back scattering improves the sensitivity of the instrument nearly 50 fold over previous versions of dynamic and static light scattering instruments, in which scattered light is collected at an angle of 90◦ . Weakly scattering systems and highly absorbing systems such as proteins, micelles, microemulsions, nanosols, dyes, pigments, and toners can be easily studied, which is not possible with traditional dynamic scattering techniques. For weakly scattering systems and highly absorbing systems the difficulty arises from the reduction in the intensity of scattered light using a scattering angle of 90◦ . This is for two reasons, the absorption reduces the power of the incident beam and the amount of scattered light is further reduced by having to pass through the sample before being detected. The HPPS instrument is outfitted with an automated optimization of the cell position to change the path length of the radiation to handle lower and higher concentration samples. The information obtainable is hydrodynamic radius, particle distribution and dispersity, solution composition, absolute molecular weight, second virial coefficient, macromolecular conformation, and shape estimation. A further advantage of this instrumentation is that data collection requires only minutes, facilitating the in situ observation of particle formation and growth. In case of multimodal distribution, resolution of two different size distribution peaks becomes difficult if they are very close to each other, resolution becomes apparent if they are at least 10 nm apart in size. Here we report the application of HPPS in monitoring the kinetics of metal cluster growth and decay as well as in determining particle sizes of various nanoscale materials. The systems were chosen to examine the applicability of HPPS to nanoresearch over a broad range of sizes. A polymer sample and Pd colloid were chosen to determine the accuracy of HPPS at 200 and 100 nm, respectively. Gold and platinum nanoparticle formation and silver nanoparticle oxidation were studied to determine the applicability to kinetics and provide a direct comparison to well established literature results. Finally, the lower limit of the instrument was examined by using well-characterized polyoxoanion samples whose size has been previously established by single crystal X-ray structures. It is our hope that establishing the applicability of the HPPS system to the study of nanoscale materials will provide researchers with a tool that will facilitate the characterization of nanoscale materials and aid in development of the synthetic methods.
460
L. D’Souza et al. / Journal of Colloid and Interface Science 279 (2004) 458–463
2. Experimental 2.1. Chemicals Duke Scientific Corporation supplied the polymer sample of polystyrene latex hydrosol with a previously established particle radius of 100 nm. Micropore water of resistance 18.2 M cm was used throughout the experiments and the apparatus was supplied by Purelab Plus, Germany. Pd(II) acetate, disodium citrate, AgNO3 , H2 PtCl6 , HAuCl4 , and acetonitrile were supplied by Fluka Chemical Corporation. Methanol and sodium borohydride were supplied by AppliChem, GmbH. KBr was supplied by Aldrich. Phosphotungstate with the well-known Keggin structure [PW11 O39 ]7− was supplied by Prof. Kortz, IUB, Germany. The almost spherical, monolacunary [PW11 O39 ]7− has a radius of 0.52 nm based on crystallographic data [26]. The gases H2 , N2 , and Ar were supplied by Messer-Griessheim, Germany and of 5.0 purity. 2.2. Apparatus and measurement description UV–visible spectra were recorded using a Varian CARY/ 100 Bio-Spectro-photometer. The samples were in the form of sol and absorbance was measured in a quartz cuvette. The scan rate applied was 100 nm min−1 . Dynamic light scattering (DLS) measurements were carried out using a high performance particle sizer supplied by Malvern Instruments and used for particle size measurement. ALV correlator V 3.0 version software was used to correlate collected scattered signals. With the exception of the kinetic measurements, 5 runs each of 30 s duration were done over the full range of the instrument and the average results have been presented. There is no difference in the sampling ranges for the measurement for any sample. The resultant peak was zoomed for the clarity of presenting the obtained results. Scanning electron microscopy (SEM) measurements have been done using JEOL JSM/5900 instrument. A drop of colloidal sol was put on the aluminum foil and the solvent allowed to evaporate. The resulting sample was mounted on the instrument for SEM measurement. All the measurements were performed at room temperature unless stated separately.
2.2.3. Preparation of silver colloids Typically 1 ml of (aminopropyl)trimethoxysilane [APS] was dissolved in 7 ml of absolute methanol, followed by the addition of 0.25 ml of 0.01 M aqueous AgNO3 solution and sonicated for 15 min. The contents were reduced by the addition of 0.05 ml of aqueous 1% NaBH4 . The reaction was carried out under ambient conditions. 2.2.4. Preparation of palladium colloids Pd nanoclusters were prepared as follows, typically 2.2 mg of Pd(II) acetate was dissolved in 10 ml of acetone, the contents were sonicated for 5 min. To the optically clear solution, 60 µl of formic acid was added and sonication was continued for 30 min, at 40 ◦ C. Thus, formed colloids were found to be stable for more than six months.
3. Results and discussion To explore whether this technique was applicable to nanoscale phenomena, the formation kinetics of gold nanoclusters was followed by HPPS. Gold was chosen because the kinetics of colloid formation using UV–visible spectroscopy is well established and can provide us with additional evidence regarding the reliability of the HPPS technique. The reaction was followed by both UV–visible spectroscopy and HPPS instrument simultaneously. Fig. 1 shows the results obtained by HPPS and UV–vis, respectively. The gold nanocluster surface plasmon lies around 525–550 nm depending upon their particle diameter [27–31]. The gradual red shift of the gold plasmon peak and particle size as measured by HPPS are in good agreement with each other. It is to be noted that there is a slight difference in the plateaus of Figs. 1a and 1c. This could be because the variation of
2.2.1. Preparation of gold colloids The procedure for colloid preparation was typically to mix 10 ml of aqueous 0.2 mM HAuCl4 with 0.5 ml of 1% sodium citrate solution under constant stirring at room temperature and an inert atmosphere. 2.2.2. Preparation of platinum colloids Platinum colloids were prepared as follows; typically 1 ml of 0.01 M H2 PtCl6 solution and 1 ml of 0.1 M TMACl solution were mixed with 25 ml of water. After stirring for 10 min under an inert atmosphere, metal ions were reduced using 1 ml of a 1% freshly prepared aqueous solution of NaBH4 .
Fig. 1. Growth kinetics of citrate stabilized gold nanoparticles. The figure shows particle size d (a), count rate (b) as measured by HPPS and peak maxima (c), absorbance (d) as measured by UV–vis spectroscopy. The red shift of peak maxima and particle size growth; absorbance value and count rate are complementary. The trend of exponential growth of nanoparticles are in good agreement with literature data [32].
L. D’Souza et al. / Journal of Colloid and Interface Science 279 (2004) 458–463
461
Fig. 2. Absorbance for platinum nanoparticles formation as measured by UV–visible spectroscopy (a) and count rate (b), particle diameter d nm (c) for platinum nanoparticles formation as obtained by HPPS. Platinum nanoparticles do not have any sharp plasmon peak. An induction period of 4 min was observed. The particles of diameter 23 nm were formed in approximately 35 min.
particle size and plasmon peaks is not consistent. The absorbance value of UV–vis spectra is found to be proportional to the count rate as measured by HPPS. The results closely match with literature reported results by Turkevich [32]. On page 66 he states that ‘the nucleation curve of gold colloids by sodium citrate reduction method contains four regions, an induction period, followed by the rapid rise at the beginning of the nucleation, a linear portion and finally a decay portion, the general nature of the curve is the characteristic of autocatalytic reaction.’ The growth process follows an exponential law of growth. It was also stated that the duration of the induction period decreases with increased temperature. An induction period is the time necessary to form an amount of acetone dicarboxylate ions necessary for nucleation. An induction period of approximately 32 min was obtained with the present experimental conditions which match with Turkevich’s reported work. The kinetics of Pt particle formation as followed by HPPS and UV–vis spectroscopy is shown in Fig. 2. The kinetics of reduction was followed using both UV–visible spectroscopy and HPPS simultaneously. The absorbance of the platinum colloid was measured at 500 nm since platinum nanoparticles do not have any specific characteristic peak in the visible range of the spectrum but they absorb in increasing amounts from the red to violet region of the spectrum [33]. The platinum ions took about 35 min for complete reaction and particle growth stops at 23 nm with the present reaction conditions (Fig. 2c). The absorbance data (Fig. 2a) and count rate (Fig. 2b) are again in good agreement with each other. Fig. 3 shows the oxidation decay kinetics of a silver system. Oxidation of Ag (0) nanoparticles to Ag (I) has been studied using HPPS and UV–visible spectroscopy. Silver colloids were prepared according to the published procedure by Lev and co-workers [34]. The silver nanoparticles form almost immediately during the reduction and within 10–20 s begin to oxidize back to the Ag-amine complex. The final particle size was 14 nm as reported by Lev (Fig. 3c) [34]. During the initial stage of reduction very large Ag particles form and then degrade very quickly within 5 min as supported by both their plasmon shift and HPPS results (Fig. 3a). The kinetics results obtained with the two instruments are again in excellent agreement and correlate each other (Figs. 3a–3d).
Fig. 3. Oxidation kinetics of Ag(0) to Ag(I). The figure shows peak maxima (a), absorbance of Ag (b) as measured by UV–visible spectroscopy and particle diameter d (c), count rate (d) as obtained by HPPS. The initially formed silver nanoparticles of size ≈ 200 nm degrade to 14 nm in about 50 min. Ag nanoparticles oxidize to Ag-amine complex.
To assess the applicability to nonmetallic substances and the upper size range of nanoscale materials, HPPS was applied to a standard sample of polymer colloidal hydrosols. Polystyrene latex hydrosol with a previously established particle diameter of 200 nm was used for testing. SEM images of polystyrene latex (not shown here) show that the particles are uniform in size with a spherical shape and the average size of the particles is 200 nm in diameter. Typically, 20 µl of polymer concentrate was diluted in 2 ml of water for HPPS studies. Fig. 4 shows the results obtained by HPPS, clearly demonstrating the particle dimension is very similar to that of SEM images, i.e., 200 nm in diameter. The applicability of the HPPS to metal clusters was then tested by examining palladium colloids. Scanning electron microscopy images were obtained using a JEOL/JSM 5900 instrument. Images of Pd clusters show that the particles are uniform in size with a spherical shape, and the average diameter of the particles is ≈ 90 nm as shown in Figs. 5a and 5b. The prepared colloid was subjected to HPPS measurement with five scans of 30 s duration. Fig. 6 shows the particle size distribution of the Pd colloidal solution obtained by HPPS measurement and found that mean particle size to be 90 nm exactly similar to that of obtained with SEM images.
462
L. D’Souza et al. / Journal of Colloid and Interface Science 279 (2004) 458–463
Fig. 6. Particle size distribution of palladium colloid as obtained from HPPS. Particle size is 90 nm. Error bars are included. The HPPS results are consistent with SEM results.
Fig. 4. Particle size distribution of the polymer hydrosol as obtained from SEM (as measured for 150 particles) (a) and HPPS (b). The mean diameter and standard deviation of particles sizes are 217 and 20 (for SEM) measurements and 212 and 10 (for HPPS), respectively.
Fig. 5. SEM image (a) and particle size distribution of palladium colloid (as measured for 150 particles) (b). The mean diameter and standard deviation of particles sizes are 91 and 8, respectively. Error bars are included.
HPPS was further used to study polyoxoanions in order to establish the lower size limit observable with this instrumentation. Two different phosphotungstates with the well-known Keggin [PW11 O39 ]7− and Wells–Dawson [P2 W18 O62 ]6− structures were synthesized according to published procedures and isolated as tetra-n-butylammonium (TBA) salts [26,35–37]. The purity of (TBA)4 H3 [PW11 O39 ] and (TBA)6 [P2 W18 O62 ] was established by FTIR and NMR studies. The almost spherical, monolacunary [PW11 O39 ]7− has a diameter of 1.04 nm and the football shaped [P2 W18 O62 ]6− has an equatorial diameter of 1.02 nm and a longitudinal diameter of 1.38 nm based on crystallographic data. Typically, a 1 mM solution of polyoxianions were prepared in Analytical Reagent grade acetonitrile and filtered through 0.4 µm filter to remove any possible dust contamination. All measurements were performed at room temperature. The hydrodynamic diameter obtained by HPPS (not shown in the figure) for (TBA)4 H3 (PW11 O39 ) is 1.02 nm and for (TBA)6 (P2 W18 O62 ) is 1.34 nm which correlate exactly with the cluster dimensions of the respective polyoxoanions. In the case of (TBA)6 (P2 W18 O62 ) the peak maxima seems to be that of the diameter of the longer axis. In both cases, the size obtained looks like that of only the polyanion unit and contributions from the mobile counter ion Bu4 N+ unit was not observed most probably, because the detection of Bu4 N+ unit is beyond the range of the instrument. These results, by distinguishing particles having very small size differences demonstrate the accuracy of HPPS at the lower limit. In conclusion, it has been demonstrated that HPPS is an extremely valuable instrumentation for nanoscale research. In particular, it offers the preparative chemist a means by which to follow particle formation and growth in situ. The duration of measurements is very short (≈ 5–30 s) as compared to other techniques such as TEM, XRD, and SEM (≈ 2–24 h). The instrumental approach is economical, much more so than TEM, XRD, or SEM. HPPS has been proven effective for studying polyoxoanion, metal colloids and polymeric samples. The kinetics of Au and Pt metal cluster growth and Ag nanoparticle oxidation has been effectively followed with HPPS. The applicability of this
L. D’Souza et al. / Journal of Colloid and Interface Science 279 (2004) 458–463
instrument to samples of 1–200 nm in diameter provides researchers working in the nanoscale regime a valuable tool for observing in situ phenomena. The capability to handle multimodal and polydispersive samples is an additional feature of this technique. If the multimodal distribution is very broad, then resolution of different particles becomes difficult since scattering intensity of particles increases with the order of 106 and hence bigger particles block the smaller one. The refractive index of the solvent plays a main role and suitable precautions should be taken to insure accurate results. Acknowledgments The authors thank Professor U. Kortz, Department of Chemistry, IUB for supplying the polyoxometalate samples, and Dr. T. Balster, Department of Physics, IUB for SEM measurements. Thanks are also due to IUB for Financial aid. L.D. is grateful to IUB for Research Scholarship. References [1] H. Bönnemann, R.M. Richards, Eur. J. Inorg. Chem. 83 (2001) 1. [2] D.L. Feldheim, C.A. Foss, Metal Nanoparticles, Synthesis, Characterization and Application, Dekker, New York, 2002. [3] H. Bönnemann, G. Braun, W. Brijoux, R. Brinkmann, K. Schulze Tilling, K. Seevogel, K. Siepen, J. Organomet. Chem. 520 (1996) 143. [4] K.J. Klabunde, Nanostructured Materials in Chemistry, Wiley–VCH, New York, 2001. [5] L. D’Souza, S. Sampath, Langmuir 16 (2000) 8510. [6] G. Viau, R. Brayner, L. Poul, N. Chakroune, E. Lacaze, F. FievetVincent, F. Fievet, Chem. Mater. 15 (2003) 486. [7] C. Pan, K. Peler, K. Philippot, B. Chaudret, F. Dassenoy, P. Lecante, M.-J. Casanove, J. Am. Chem. Soc. 123 (2001) 7584. [8] R. Richards, W. Li, S. Decker, C. Davidson, O. Koper, V. Zaikovski, A. Volodin, T. Rieker, K.J. Klabunde, J. Am. Chem. Soc. 122 (2000) 4921. [9] H. Bönnemann, W. Brijoux, R. Brinkmann, E. Dinjus, T. Joussen, B. Korall, Angew. Chem. Int. Ed. 30 (1991) 1344. [10] M. Watzky, R.G. Finke, J. Am. Chem. Soc. 119 (1997) 10382.
463
[11] J.A. Widegren, J.D. Aiken, S. Özkar, R.G. Finke, Chem. Mater. 13 (2001) 312. [12] A. Kolmakov, D.W. Goodman, Chem. Rec. 2 (2002) 446. [13] M. Magnuson, D. Lytle, C. Frietch, C. Kelty, Anal. Chem. 73 (2001) 4815. [14] V. Bogatyrev, L. Dykman, Y. Krasnov, V. Plotnikov, N. Khlebtsov, Colloid J. 64 (2002) 671. [15] M. Ballauff, Macromol. Chem. Phys. 204 (2003) 220. [16] S. Takata, M. Shibayama, R. Sasabe, H. Kawaguchi, Polymer 44 (2003) 495. [17] K. Kostakis, S. Mourmouris, D. Charalabidis, M. Pitsikalis, Eur. Phys. J. E 10 (2003) 55. [18] S. Chen, W. Chen, F. Mallamace, Science 300 (2003) 619. [19] L. Bes, S. Angot, A. Limer, D. Haddleton, Macromolecules 36 (2003) 2493. [20] I. Michalska, A. Krysztafkiewicz, M. Bogacki, T. Jesionowski, J. Chem. Technol. Biotechnol. 78 (2003) 452. [21] K. Mogyorosi, I. Dekany, J. Fendler, Langmuir 19 (2003) 2938. [22] G. Artioli, R. Grizzetti, L. Carotenuto, C. Piccolo, C. Colella, B. Liguori, R. Aiello, P. Frontera, Stud. Surf. Sci. Catal. 142 (2002) 45. [23] M. Bellour, A. Knaebel, J. Harden, F. Lequeux, J. Munch, Phys. Rev. E 67 (2003) 031405. [24] J. Gregory, Nanoparticles in Solids and Solutions, Kluwer Academic, Dordrecht, 1996. [25] J.P. Wilcoxon, J.E. Marin, D.W. Schaefer, Phys. Rev. Lett. 58 (1987) 1051. [26] J.F. Keggin, Nature 131 (1933) 908. [27] M.-C. Daniel, D. Astruc, Chem. Rev. 104 (2004) 293. [28] S. Link, Z.L. Wang, M.A. El-Sayed, J. Phys. Chem. B 103 (1999) 3529. [29] M.B. Mohamed, Z.L. Wang, M.A. El-Sayed, J. Phys. Chem. A 103 (1999) 10255. [30] E.C.H. Sykes, F.J. Williams, M.S. Tikhov, L.S. Richard, J. Phys. Chem. B (2001). [31] R.N. Jana, L. Gearheart, J.C. Murphy, Chem. Mater. 13 (2001) 2313. [32] J. Turkevich, P.C. Stevenson, J. Hillier, Discuss. Faraday Soc. 11 (1951) 55. [33] A.J. Creighton, D.G. Eadon, J. Chem. Soc. Faraday Trans. 87 (1991) 3881. [34] S. Bharathi, N. Fishelson, O. Lev, Langmuir 16 (1999) 1929. [35] R. Contant, Inorg. Synth. 27 (1990) 105. [36] R. Contant, Can. J. Chem. 65 (1987) 568. [37] B. Dawson, Acta Crystallogr. 6 (1953) 113.
